Two-dimensional nuclear magnetic resonance spectroscopy has been used to monitor proton-deuterium exchange rates (k(obs)) for more than 30 residues in turkey ovomucoid third domain. To test whether exchange is governed by global unfolding, rates were measured over a wide range of pH and temperatures where the change in the free energy of unfolding (Delta G(u) degrees) is known [Swint, L., & Robertson, A. D. (1993) Protein Sci. 2, 2037-2049; Swint-Kruse, L., & Robertson, A. D. (1995) Biochemistry 34, 4724-4732]. Under conditions where EX2 kinetics are observed, a subset of 6_11 residues exhibits a one-to-one correlation with global stability. These residues are all located in central regions of secondary structures. Many other sites show varied degrees of correlation with Delta G(u) degrees, while some are slower than expected on the basis of Delta G(u) degrees alone. Preliminary evidence suggests that the latter is due to deviation from EX2 kinetics, even though experimental conditions are relatively mild (pH* 3 and 40 degrees C) compared to those in which deviations were observed for bovine pancreatic trypsin inhibitor. These results, together with similar observations for hen egg white lysozyme and barnase, suggest that EX2 kinetics should not be assumed when interpreting exchange studies.